use TeO, clear

// Standardize outcomes
foreach x in harass_school harass_work harass_university ///
		harass_store harass_street harass_transport harass_police harass_hospital harass_bank ///
		harass_pubserv harass_other {
	qui sum `x'
	local `x'm=r(mean)
	sum `x' 
	local sd=r(sd)
	gen `x'st=(`x'-``x'm')/`sd'  
}

//Regressions
local sample if bpl=="1101" &female==1
local inter interm 
local controls i.birthyear i.religion tmuslim  

estimates clear

reg harass_schoolst `controls' `inter' `sample', cl(religion)
eststo m1

reg harass_workst `controls' `inter' `sample', cl(religion)
eststo m2

reg harass_universityst `controls' `inter' `sample', cl(religion)
eststo m3

reg harass_storest `controls' `inter' `sample', cl(religion)
eststo m4

reg harass_streetst `controls' `inter' `sample', cl(religion)
eststo m5

reg harass_transportst `controls' `inter' `sample', cl(religion)
eststo m6

reg harass_policest `controls' `inter' `sample', cl(religion)
eststo m7

reg harass_hospitalst `controls' `inter' `sample', cl(religion)
eststo m8

reg harass_bankst `controls' `inter' `sample', cl(religion)
eststo m9

reg harass_pubservst `controls' `inter' `sample', cl(religion)
eststo m10

reg harass_otherst `controls' `inter' `sample', cl(religion)
eststo m11

esttab m* using "Table3.csv", star(+ 0.1 * 0.05 ** 0.01 *** 0.001) replace ///
		cells(b(fmt(a3) star) se(par)) stats(N r2)  ///
		keep(interm) 
		
